The depth-3 circuit model has emerged as a key step in understanding general circuits. In this talk we will focus on the question of testing whether a given *multilinear* depth-3 circuit is zero. In particular, we give nontrivial constructions of hitting-sets for *low-distance* multilinear depth-3. The main phenomena here is "low-support rank concentration"; naturally, we conjecture that it can be attained for general depth-3 circuits as well.